Digital Restoration of Mural Images using TSCPMA with High Frequency Generation Scheme
Sujata G. Bhele, Shashank Shriramwar, Poonam T. Agarkar · 2023
State of art techniques based on patch matching have been developed to inpaint degraded regions of mural images. The application of these developed techniques are limited and are not versatile to cope with geographical noises. The performance of the pioneer algorithm developed by Criminisi suffers when order of the pixel chosen for patch matching is not appropriate and consumes long time. We proposed an efficient texture-structure conserving patch matching algorithm (TSCPMA) with state of art method developed by Purkait and Chanda, 2002 [1] for high frequency image enhancing. The image enhancement approach involves high frequency generation scheme using TSCPMA and anisotropic diffusion using Beltrami filtering. Though this technique achieves remarkable results in variety of applications, we have introduced the TSCPMA scheme to high textured image and the complete algorithm to mural images. The algorithm have ability to improve the visual quality of an image by strengthening the edges while preserving other smooth regions. While the TSCPMA reduces the time complexity in patch matching mechanism, the experimental analysis showed accurate recovery of high quality textures and structures.